Module 6: Deadlocks
Reading: Chapter 7
Objective:
To develop a description of deadlocks, which
prevent sets of concurrent processes from
completing their tasks
To present a number of different methods for
preventing or avoiding deadlocks in a
computer system.
1
Resource owned by other
waiting process
System Process
Model waits on
resource
Dead
lock
Char Deadlocks Pr oblem
acter
i zatio
n
Handling
Deadlocks
2
The Deadlock Problem
A set of blocked processes each holding a
resource and waiting to acquire a resource held
by another process in the set.
Recall the example with semaphores
semaphores A and B, initialized to 1
P0 P1
wait (A); wait(B)
wait (B); wait(A)
Example – law passed by Kansas Legislature
“When two trains approach each other at a crossing,
both shall come to a full stop and neither shall start
up again until the other has gone.”
3
Instances Resources
Resource owned by other
Use
waiting process
s
Request
ation
Release
System Process
Oper
Model waits on
resource
Dead
lock
Char Deadlocks Pr oblem
acter
i zatio
n
Handling
Deadlocks
4
System Model
Resource types R1, R2, . . ., Rm
CPU cycles, memory space, I/O devices
Each resource type Ri has Wi instances.
2 printers, three hard drives, etc.
Each process utilizes a resource as follows (using
system calls):
request
use
release
Deadlock example
Three processes each hold a CD drive.
Each process requires a 2nd drive.
Multithreaded programs are good candidates for
deadlock
Threads share many resources.
5
Instances Resources
Resource owned by other
Use
s waiting process
Request
ation
Release
System Process
Oper
Model waits on
resource
Dead
lock
Char Deadlocks Pr oblem
acter
i zatio
n
tio y
di sar
Mutual Handling
ns
on s
C ece
Exclusion
Deadlocks
N
No Pre-
Hold and
emption Circular
Wait
Wait
Resources Assignment
Resource- Request
Processes
Allocation Edge
Vertex
6Graph
Deadlock Characterization
Deadlock can arise if four conditions hold simultaneously.
Mutual exclusion: only one process at a time can use a
resource.
Hold and wait: a process holding at least one resource is
waiting to acquire additional resources held by other
processes.
No preemption: a resource can be released only
voluntarily by the process holding it, after that process
has completed its task.
Circular wait: there exists a set {P0, P1, …, Pn} of waiting
processes such that P0 is waiting for a resource that is
held by P1, P1 is waiting for a resource that is held by
P2, …, Pn–1 is waiting for a resource that is held by
Pn, and Pn is waiting for a resource that is held by P0.
• When the first 3 conditions exist, the circular wait is an
indication of a deadlock.
• The first 3 conditions do not imply necessarily a deadlock,
since the circular wait may not occur.
7
Resource-Allocation Graph
A set of vertices V and a set of edges E.
V is partitioned into two types:
P = {P1, P2, …, Pn}, the set consisting of
all the processes in the system.
R = {R1, R2, …, Rm}, the set consisting
of all resource types in the system.
Edges
request edge – directed edge Pi Rj
assignment edge – directed edge Rj
Pi
8
Resource-Allocation Graph (Cont.)
Process
Resource Type with 4 instances
Pi requests instance of Rj
Pi
Rj
Pi is holding an instance of Rj
Pi
Rj
9
Example of a Resource Allocation Graph
Is there deadlock?
10
Using the graphs
Assume that the first three conditions for
deadlock exist.
Mutual Excl., hold and wait, no pre-emption
To show a deadlock, must show that no
circular wait exists.
A process can terminate without any
waiting on another, and then the others.
<P3, P2, P1> is a termination sequence: all
processes can terminate in this order.
11
Resource Allocation Graph With A Deadlock
Circular waits:
P1 R1 P2 R3 P3
R2 P1
P2 R3 P3 R2 P2
No proc. can terminate
No possible way out.
12
Graph with a circular wait, but not deadlock (why?)
13
Basic Facts
If graph contains no cycles no deadlock.
If graph contains a cycle
if only one instance per resource type, then
deadlock.
if several instances per resource type,
possibility of deadlock.
• Must determine if a process can terminate
and if so, will it allow other processes to
terminate.
14
Instances Resources
Resource owned by other
Use
waiting process
s
Request
ation
Release
System Process
Oper
Model waits on
resource
Dead
lock
Char Deadlocks Pr oblem
acte rizat
r y on i Ignore the
ssa ons Mutual Handling problem
e i
ec dit Exclusion Avoid Recovery
N on Deadlocks
C No Pre- Prevent
Hold and Detection
emption Circular
Wait
Wait No With
deadlocks deadlocks
Resources Assignment
Resource- Request
Processes
Allocation Edge
Vertex
Graph
15
Methods for handling deadlocks
Do not enter deadlock state
Deadlock prevention
• Disallow 1 of the 4 necessary conditions of deadlock
occurrence.
• Difficult and restraining – can lead to low device
utilization and low throughput.
Deadlock avoidance
• Do not grant a resource request if this allocation might lead
to deadlock.
Allow deadlocks
Periodically check for the presence of deadlock and then
recover from it.
Ignore the problem:
Ignore the problem and pretend that deadlocks never
occur in the system.
Used by most operating systems, including
UNIX/Windows.
16
Deadlock Prevention: prevent at
Ignore the
Handling problem
Avoid Recovery
Deadlocks
least one of the four conditions that lead to
Prevent Detection
No With
deadlock
deadlocks deadlocks
Mutual exclusion: reduce as much as possible the
use of shared resources and critical sections
(almost impossible).
Hold and wait: a process that requests new
resources cannot hold up other processes (ask for
all resources at once).
No pre-emption: If a process asks for resources
and cannot obtain them, it is suspended and its
resources already held are released.
Circular wait: define an request ordering for
resources, a process must ask for resources in
this order (e.g. always ask for the printer before
the tape drive) (see section 7.4.4 in the textbook
for more details).
17
Ignore the
Handling problem
Avoid Recovery
Deadlock Avoidance
Deadlocks
Prevent Detection
No With
deadlocks deadlocks
Requires that the system has some additional a priori
information available.
Simplest and most useful model requires that each
process declare the maximum number of
resources of each type that it may need.
The deadlock-avoidance algorithm dynamically
examines the resource-allocation state to ensure
that there can never be a circular-wait condition.
Resource-allocation state is defined by the
number of available and allocated resources, and
the maximum demands of the processes.
18
Safe State
When a process requests an available
resource, system must decide if
immediate allocation leaves the system
in a safe state.
What is a safe state?!
System is in safe state if there exists a
safe sequence of all processes.
19
Basic Facts
If a system is in safe state no
deadlock possible.
If a system is in unsafe state
possibility of deadlock.
Avoidance ensure that a
system will never enter an
unsafe state.
20
Safe Sequence
Safe sequence is a sequence witnessing that we can run
all processes to completion
P1 is the process that can be run to completion using
only the available resources
P2 is the process that can be completed when P1
completes and releases its resources…
Sequence <P1, P2, …, Pn> is safe if for each Pi, the
resources that Pi can still request can be satisfied by
currently available resources + resources held by all the
Pj, with j<i.
If P resource needs are not immediately available,
i
then Pi can wait until all Pj have finished.
When Pj is finished, Pi can obtain needed resources,
execute, return allocated resources, and terminate.
When Pi terminates, Pi+1 can obtain its needed
21 resources, and so on.
Resource-Allocation Graph Algorithm
First consider the simpler case of one-instance
resources
Note: in such case, a cycle in the Resource-
Allocation graph implies deadlock
Lets introduce claim edges:
Claim edge Pi Rj indicates that process Pj may
request resource Rj; represented by a dashed line.
Claim edge is converted to request edge when a
process requests a resource.
When a resource is released by a process,
assignment edge reconverts to a claim edge.
Resources must be claimed a priori in the system.
The algorithm: if satisfying request creates a cycle
in the modified R-A graph (including claim edges),
22 reject the request
Resource-Allocation Graph with Claim
Edges
23
Unsafe State In Resource-Allocation Graph
24
Banker’s Algorithm
For the case when resources have multiple
instances.
Again, each process must a priori claim the
resources it is going to use
Must also specify quantity
When a process requests a resource it may have
to wait, even if the resource is currently available
The banker’s algorithm decides whether to grant the
resource
Assumes that when a process gets all resources it
wants, within finite amount of time it will finish
and release them
25
Data Structures for the Banker’s
Algorithm
Let n = number of processes, and m = number of resources
types.
Available: Vector of length m. If available[j] = k, there
are k instances of resource type Rj available.
Max: n x m matrix. If Max[i,j] = k, then process Pi
may request at most k instances of resource type Rj.
Allocation: n x m matrix. If Allocation[i,j] = k then Pi
is currently allocated k instances of Rj.
Need: n x m matrix. If Need[i,j] = k, then Pi may need
k more instances of Rj to complete its task.
Need [i,j] = Max[i,j] – Allocation [i,j].
26
Banker’s Algorithm for Process Pi
Describes how to respond to a resource request by process Pi.
Request = request vector for process Pi.
If Requesti [ j ] = k then process Pi wants k instances of
resource type Rj.
1. If Requesti Needi go to step 2. Otherwise, raise error
condition, since process has exceeded its maximum claim.
2. If Requesti Available, go to step 3. Otherwise Pi must wait,
since resources are not available.
3. Pretend to allocate requested resources to Pi by modifying the
state as follows:
Available = Available - Requesti;
Allocationi = Allocationi + Requesti;
Needi = Needi – Requesti;
If the resulting state is safe the resources are allocated to Pi.
If unsafe Pi must wait, and the old resource-allocation state is
restored
27
Safety Algorithm
But how do we test whether a state is safe?
Remember, there must be a safe sequence in a safe state.
So, we will try to find a safe sequence.
Find a process that can be satisfied with the current
resources, give it the resources and let it finish.
(include the need ressources).
This would be our P1.
When P1 finishes, release its resources, and repeat
How long should repeat?
Until all processes are finished – then the state was
indeed safe.
Or until the we cannot find a process to finish – then the
state was unsafe.
Remember, this is all as-if.
28
Safety Algorithm
1. Let Work and Finish be vectors of length m and n,
respectively. Initialize:
Work = Available
Finish [i] = false for i = 0,1, 2, 3, …, n-1.
2. Find and i such that both:
(a) Finish [ i ] = false // not yet finished process
(b) Needi Work // that can be finished with available
resources
If no such i exists, go to step 4.
3. Work = Work + Allocationi
Finish[ i ] = true
go to step 2.
4. If Finish [ i ] == true for all i, then the initial state was
safe,
Otherwise, it was unsafe
29
Safety Algorithm
Great! Now I understand how it works.
Too bad I don’t understand why is works.
We have tried only one way to find a safe
sequence
If we failed, a different choice of the satisfiable
processes might have lead to a safe sequence,
but we just made bad choices leading us to a
dead end
That is actually not true
If we found a satisfiable process, let it run to
completion and then released its resources, we
are better off (more resources available) then how
we started
So, it does not harm us to choose any satisfiable
process
30
Example of Banker’s Algorithm
5 processes P0 through P4; 3 resource types A (10
instances), B (5 instances), and C (7 instances).
Snapshot at time T0:
Allocation Max Available Need
ABC ABC ABC ABC
P0 010 753 ??? 743
P1 2 0 0 322 122
P2 3 0 2 902 600
P3 2 1 1 222 ???
P4 002 433 ???
Is the system safe?
31
Example: P1 requests (1,0,2)
Check that Request Need[1](i.e, (1,0,2) (1,2,2)) true.
Check that Request Available (i.e, (1,0,2) (3,3,2)) true.
Allocation Need Available
ABC ABC ABC
P0 0 1 0 743 230
P1 302 020
P2 301 600
P3 211 011
P4 002 431
Executing safety algorithm shows that sequence <P1, P3, P4,
P0, P2> satisfies safety requirement.
Can request for (3,3,0) by P4 be granted?
Can request for (0,2,0) by P0 be granted?
32
Ignore the
Handling problem
Deadlock Detection
Avoid Recovery
Deadlocks
Prevent Detection
No With
deadlocks deadlocks
Resource access are granted to processes
whenever possible. The OS needs:
an algorithm to check if deadlock is present
an algorithm to recover from deadlock
The deadlock check can be performed at
every resource request
Such frequent checks consume CPU time
33
Single Instance of Each Resource
Type
Maintain wait-for graph
Nodes are processes.
Pi Pj if Pi is waiting for Pj.
Periodically invoke an algorithm that searches for
a cycle in the graph.
Requires an order of n2 operations, where n is the
number of vertices in the graph.
Don’t want to run it too often
34
Resource-Allocation Graph and Wait-for
Graph
Resource-Allocation Graph Corresponding wait-for graph
35
Several Instances of a Resource Type
Available: A vector of length m indicates the
number of available resources of each type.
Allocation: An n x m matrix defines the number
of resources of each type currently allocated to
each process.
Request: An n x m matrix indicates the current
request of each process. If Request [ij] = k, then
process Pi is requesting k more instances of
resource type. Rj.
36
Detection Algorithm
Idea:
Reuse the ideas from the Safety algorithm
Essentially, check whether it is possible from the
current state to finish all processes
In other words, try to construct a safe sequence
if succeeded, not it deadlock (at least not yet)
if failed (some processes have not been finished),
these processes are deadlocked
The processes that do not hold resources are
considered “finished” (they are not blocking
others, so they cannot participate in a deadlock
cycle)
37
Detection Algorithm
1. Let Work and Finish be vectors of length m and n,
respectively Initialize:
(a) Work = Available // as-if, remember
(b) For i = 1,2, …, n,
if Allocationi 0 or Requesti 0
then Finish[i] = false;
otherwise, Finish[i] = true
2. Find an index i such that both:
(a) Finish[i] == false
(b) Requesti Work
If no such i exists, go to step 4.
38
Detection Algorithm (Cont.)
3. Work = Work + Allocationi
Finish[i] = true
go to step 2.
4. If Finish[i] == false, for some i, 1 i n,
then the system is in deadlock state.
Moreover, if Finish[i] == false, then Pi is
deadlocked.
Algorithm requires an order of O(m x n2) operations to
detect whether the system is in deadlocked state.
39
Example of Detection Algorithm
Five processes P0 through P4; three resource types
A (7 instances), B (2 instances), and C (6 instances).
Snapshot at time T0:
Allocation Request Available
ABC ABC ABC
P0 0 1 0 000 000
P1 200 202
P2 303 000
P3 211 100
P4 002 002
Sequence <P0, P2, P3, P1, P4> will result in Finish[i] = true for
all i.
40
Example (Cont.)
P2 requests an additional instance of type C.
Request
ABC
P0 0 0 0
P1 201
P2 001
P3 100
P4 002
State of system?
Can reclaim resources held by process P , but
0
insufficient resources to fulfill other processes’ requests.
Deadlock exists, consisting of processes P , P , P , and
1 2 3
P4.
41
Detection-Algorithm Usage
When, and how often, to invoke depends
on:
How often a deadlock is likely to occur?
How many processes will be affected by the
deadlock when detected?
If detection algorithm is invoked arbitrarily,
there may be many cycles in the resource
graph
Difficult to tell which of the many
deadlocked processes “caused” the
42 deadlock.
Ignore the
Handling
Recovery from Deadlock:
problem
Avoid Recovery
Deadlocks
Prevent Detection
Process Termination No
deadlocks
With
deadlocks
Abort all deadlocked processes.
Abort one process at a time until the deadlock
cycle is eliminated.
In which order should we choose to abort?
Priority of the process.
How long process has computed, and how much
longer to completion.
Resources the process has used.
Resources process needs to complete.
How many processes will need to be terminated.
Is process interactive or batch?
43
Recovery from Deadlock: Resource
Preemption
Selecting a victim – minimize cost.
Rollback – return to some safe state,
restart process for that state.
Need to create and maintain checkpoints
for rolling back to safe points
Starvation – same process may always be
picked as victim,
Include number of rollback in cost factor.
44
Instances Resources
Resource owned by other
Use
waiting process
s
Request
ation
Release
System Process
Oper
Model waits on
resource
Dead
lock
Char Deadlocks Pr oblem
acte rizat
r y on i Ignore the
ssa ons Mutual Handling problem
e i
ec dit Exclusion Avoid Recovery
N on Deadlocks
C No Pre- Prevent
Hold and Detection
emption Circular
Wait
Wait No With
deadlocks deadlocks
Resource- Vertex Edge
Ressources Request Assignment
Allocation Processes
Graph
45
